GeoLife: Dataset from the GeoLife project with daily routine trajectories of 182 users, collected between 2007 to 2012, and with the information of transportation modes of 72 users. This dataset has a total of 9,765 labeled trajectories, with a total of 5,506,967 points, and 11 classes of transportation modes: car, train, subway, taxi, bus, walk, bike, boat, airplane, run and motorcycle. In this dataset the sampling rate is high, which means that the points were collected with seconds of difference between them, leading this very dense dataset. The original dataset has some outliers and issues that need to be fixed for the extraction of numerical features required by some techniques, which lead to the necessity of preprocessing the dataset. The preprocessing consisted on: (i) removing duplicated records, (ii) splitting the trajectories where two consecutive points had more than 300 seconds of difference between them, as it represents a large gap in such dense dataset, (iii) removing trajectories with less than 100 points, as it represents a very small portion of time due to the density, (iv) excluding the transportation modes with too less trajectory examples, which are the classes of airplane, boat, running and motorcycle, (v) removing trajectories with unreal average velocity given the transportation mode, as for instance, trajectories labeled with walk with average velocity of more than 10m/s, and, finally, from the remaining trajectories, we (vi) selected a proportion of 20% of the trajectories of each transportation mode, as some techniques were unable to perform in a reasonable amount time. The resultant dataset has a total of 1,763 trajectories, with trajectories varying from 100 to 12,224 points.